This book presents Error-Selective Learning (ESL), an error-driven mo-del of OT phonological acquisition in which is both restrictive and gradual. Together, these two properties provide a model that can derive many attested intermediate stages in phonological develop-ment, while also understanding how learners eventually converge on the target grammar. ESL is restrictive because it relies on a restrictive ranking algorithm (Prince and Tesar 2004; Hayes 2004) to learn from its errors. At the same time ESL is gradual, because it uses a novel method for selecting errors that drive minimal changes to the grammar from each developmental stage to the next. The book synthesizes the existing arguments for BCD algorithms, presents the mechanics and rationale behind Error-Selective Learning, and provides cross-linguistic support for the ESL approach via attested intermediate stages of L1 phonological development. The book’s results also emphasize the crucial roles for stored errors, throughout the slow and extended process of building a new phonology.